Upscaling and simulation of composite gas reservoirs with thin-interbedded shale and tight sandstone

• An improved upscaling and simulation method for composite gas reservoirs with interbedded shale and tight sandstone.
• We regard the two types of rocks as dual porous media to replace single porous medium used in the traditional methods.
• The primary- and secondary- hydraulic fracture are modeled as discrete fracture and continuous medium respectively.
• The method can avert the challenge of upscaling the parametric curves of different rocks.

This paper focused on the reservoir simulation of composite gas reservoirs with interbedded shale and tight sandstone. The mentioned simulation confronts a series of problems such as computational inefficiency and serious convergence issues due to the numerous extremely thin layers in such reservoirs. Although various upscaling methods can solve these problems, they only reflect the average flow. Thus, we present an improved upscaling and simulation method. Firstly, we regard two types of rocks as dual porous media to replace single porous media used in traditional methods. Secondly, hydraulic fracturing is considered by modeling the primary fractures as discrete fractures and idealizing the secondary fractures as continuous. As such, the total reservoir is classified into three regions: (1) the triple porosity region, which contains pores of the two rocks and the secondary fractures; (2) the dual porosity region, which contains pores of the two rocks; and (3) the discrete fracture region, which indicates the primary fracture. Thirdly, the formulas for the transfer term between any two types of porosities are derived. The derived formulas are similar to but different from the existing formula. Furthermore, the upscaling methods for porosity and permeability are presented. Notably, our method can avert the challenge of upscaling the parametric curves of different rocks, e.g., capillary pressure, relative permeability, and gas adsorption–desorption. Finally, we use a real gas reservoir as an example to assess three schemes, i.e., fine grid-cells with singular porosity (FGSP), coarse grid-cells with singular porosity (CGSP), and coarse grid-cells with dual porosity (CGDP). Consequently, the CGDP results are much closer to those of FGSP, with just 1/25 grid cells and shorter computing time, unlike CGSP, which is not capable of reflecting the flow differences between the shale and siltstone. Therefore, we conclude that our method can not only improve the computing efficiency but can also guarantee accuracy.